INDEX

Explanations

sequences of numbers and symbols that resemble codes or identifiers

oai_token-act-pair · gpt-3.5-turbo

colons and numeric codes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Frie

-0.78

 Takeru

-0.77

 Yuri

-0.77

Dul

-0.75

ents

-0.74

ublic

-0.72

 Ryder

-0.72

ulent

-0.72

 secretaries

-0.71

yg

-0.70

POSITIVE LOGITS

1.34

 TAMADRA

1.21

1.20

VII

0.98

0.93

0.91

 Seven

0.89

Activations Density 0.159%

sequences of numbers and symbols that resemble codes or identifiers

colons and numeric codes

No Comments

No Known Activations