INDEX

Explanations

mentions or references to the term "op" followed by a single-digit number, possibly related to surgery or operations

oai_token-act-pair · gpt-3.5-turbo

references to "Op" or "operational" terms, likely related to technology or systems

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Lauder

-0.82

ILLE

-0.71

IGHTS

-0.62

 intending

-0.61

 Sins

-0.61

 Leviathan

-0.61

BIP

-0.60

 EntityItem

-0.59

 FANT

-0.59

Tib

-0.59

POSITIVE LOGITS

inion

1.44

yright

1.36

olitan

1.27

rompt

1.26

edia

1.21

sych

1.21

odcast

1.13

ocalypse

1.12

lication

1.12

olitics

1.09

Activations Density 0.031%

mentions or references to the term "op" followed by a single-digit number, possibly related to surgery or operations

references to "Op" or "operational" terms, likely related to technology or systems

No Comments

No Known Activations