INDEX

Explanations

years and numbers

oai_token-act-pair · gpt-3.5-turbo

instances of an end-of-text marker

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

ãĤĬ

-0.57

Thumbnail

-0.55

ãĥ¥

-0.54

ity

-0.54

cule

-0.54

¡

-0.53

ãĤ£

-0.53

minster

-0.52

ingly

-0.52

EO

-0.51

POSITIVE LOGITS

 constitutes

0.71

 including

0.70

udeb

0.70

 wherein

0.69

zbollah

0.69

 consists

0.69

 culminating

0.69

Downloadha

0.68

 comprising

0.67

 emerges

0.66

Activations Density 0.416%

years and numbers

instances of an end-of-text marker

No Comments

No Known Activations