INDEX

Explanations

measurements and numerical values expressed in millimeters

oai_token-act-pair · gpt-3.5-turbo

numerical measurements and values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

nels

-0.89

querque

-0.79

inance

-0.75

agents

-0.73

ribly

-0.72

tarians

-0.69

matter

-0.69

erville

-0.68

essee

-0.67

nel

-0.67

POSITIVE LOGITS

0.98

0.96

SourceFile

0.92

0.91

0.89

æ©Ł

0.89

0.84

 ILCS

0.83

Activations Density 0.025%

measurements and numerical values expressed in millimeters

numerical measurements and values

No Comments

No Known Activations