INDEX

Explanations

news agencies or sources mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

mentions of the AFP (Agence France-Presse) news agency

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

stood

-0.97

tons

-0.76

croft

-0.75

hematic

-0.65

 Heavenly

-0.65

coming

-0.65

atron

-0.61

order

-0.61

comes

-0.61

chest

-0.60

POSITIVE LOGITS

UTERS

1.02

AFP

0.99

DEF

0.91

aukee

0.87

 confir

0.77

uve

0.76

 PHOTO

0.76

perty

0.73

clair

0.73

AFP

0.73

Activations Density 0.005%

news agencies or sources mentioned in the text

mentions of the AFP (Agence France-Presse) news agency

No Comments

No Known Activations