INDEX

Explanations

contractions involving 'isn't' and 'doesn't'

oai_token-act-pair · gpt-3.5-turbo

negations or phrases expressing disagreement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 princ

-0.67

 symp

-0.65

 prelim

-0.63

 guid

-0.61

 nearest

-0.61

 FANT

-0.60

Introduced

-0.59

interstitial

-0.58

 passing

-0.57

 pressing

-0.57

POSITIVE LOGITS

't

1.59

Â´

0.96

ÃŃ

0.88

uts

0.86

ned

0.86

nit

0.83

ates

0.79

itates

0.78

âĤ¬

0.74

Activations Density 0.127%

contractions involving 'isn't' and 'doesn't'

negations or phrases expressing disagreement

No Comments

No Known Activations