INDEX

Explanations

the phrases where someone expresses their opinion or belief

oai_token-act-pair · gpt-3.5-turbo

phrases indicating expressions of personal opinion or assertion

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

mes

-0.74

govtrack

-0.64

ukong

-0.63

awar

-0.61

 decom

-0.60

leans

-0.60

 arra

-0.60

ahime

-0.59

Lenin

-0.59

uga

-0.56

POSITIVE LOGITS

 confidently

0.92

 anecd

0.77

quist

0.76

ħĭ

0.74

idate

0.74

 vividly

0.68

 Attribution

0.66

 unequivocally

0.66

rist

0.65

0.64

Activations Density 0.095%

the phrases where someone expresses their opinion or belief

phrases indicating expressions of personal opinion or assertion

No Comments

No Known Activations