INDEX

Explanations

entities with names or specific words in all caps

oai_token-act-pair · gpt-3.5-turbo

references to companies or corporate entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 redistributed

-0.56

"]=>

-0.55

sbm

-0.53

alist

-0.53

Article

-0.53

 airst

-0.52

 Â»

-0.50

 implicit

-0.49

 Sagan

-0.49

 Moss

-0.49

POSITIVE LOGITS

1.08

,-

0.92

KEN

0.79

)=(

0.77

,...

0.73

PIN

0.71

,.

0.69

.,

0.69

Merit

0.68

,,

0.67

Activations Density 0.381%

entities with names or specific words in all caps

references to companies or corporate entities

No Comments

No Known Activations