INDEX

Explanations

words related to valuable items

oai_token-act-pair · gpt-3.5-turbo

terms related to value and valuables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

pread

-0.78

 EntityItem

-0.70

olitan

-0.70

soever

-0.70

 Bread

-0.67

RAFT

-0.66

 Rapp

-0.65

sein

-0.65

éĹĺ

-0.63

landish

-0.63

POSITIVE LOGITS

uations

1.11

ibr

1.04

ueless

1.03

entin

0.97

val

0.95

uing

0.91

anches

0.89

ipers

0.89

enture

0.89

ancy

0.87

Activations Density 0.005%

words related to valuable items

terms related to value and valuables

No Comments

No Known Activations