INDEX

Explanations

words related to relationships between different variables or concepts

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a connection or correlation between concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

OGR

-0.79

tsky

-0.78

nell

-0.74

stal

-0.72

entimes

-0.72

kl

-0.72

fl

-0.71

di

-0.70

cca

-0.70

gn

-0.69

POSITIVE LOGITS

 sexes

0.77

 disparate

0.70

 observable

0.66

 scarce

0.63

 genders

0.62

 criminality

0.62

two

0.61

ãĤ¬

0.61

between

0.61

urst

0.59

Activations Density 0.030%

words related to relationships between different variables or concepts

phrases that indicate a connection or correlation between concepts

No Comments

No Known Activations