INDEX

Explanations

keywords related to software or technology

oai_token-act-pair · gpt-3.5-turbo

references to the term "FW" and its variations in context

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

perate

-0.65

 compounded

-0.61

ãĥŁ

-0.60

 mend

-0.59

*/(

-0.59

 istg

-0.58

 subsequent

-0.58

 Austral

-0.58

ially

-0.57

rises

-0.57

POSITIVE LOGITS

FW

1.06

fw

0.96

tsky

0.86

nesday

0.82

assium

0.81

ledged

0.79

esome

0.79

ometimes

0.79

olicy

0.78

linger

0.76

Activations Density 0.004%

keywords related to software or technology

references to the term "FW" and its variations in context

No Comments

No Known Activations