INDEX

Explanations

phrases related to holes or vulnerabilities in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases related to flaws or gaps in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

leans

-0.73

llor

-0.72

NOW

-0.71

IDE

-0.66

eas

-0.64

onym

-0.61

ALE

-0.60

killed

-0.59

Cru

-0.59

APH

-0.59

POSITIVE LOGITS

effic

1.33

efficiency

1.16

 lieu

1.11

roads

1.04

clusions

1.02

escap

1.01

 relation

0.99

between

0.99

ordinate

0.97

animate

0.93

Activations Density 0.194%

phrases related to holes or vulnerabilities in various contexts

phrases related to flaws or gaps in various contexts

No Comments

No Known Activations