INDEX

Explanations

names, specifically the name "Larry"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name 'Larry'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

minist

-1.09

-0.81

requ

-0.77

option

-0.73

hire

-0.70

words

-0.69

eworld

-0.68

notation

-0.67

alez

-0.66

repe

-0.66

POSITIVE LOGITS

 Summers

0.97

 Fitzgerald

0.90

 Ellison

0.89

 Kramer

0.89

 Hogan

0.85

 Nass

0.83

Fly

0.83

 Kaufman

0.81

 Cohen

0.81

Lur

0.80

Activations Density 0.021%

names, specifically the name "Larry"

mentions of the name 'Larry'

No Comments

No Known Activations