INDEX

Explanations

names or phrases containing the word "Jac" with varying levels of relevance indicated by activation values

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names, particularly "Jac" and "Coch."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Skydragon

-0.74

anguage

-0.72

 checkout

-0.66

wagen

-0.64

doms

-0.63

 Dull

-0.62

 Reincarn

-0.62

 partName

-0.62

 dime

-0.62

 Admir

-0.61

POSITIVE LOGITS

lyn

1.15

obal

1.14

xon

1.01

iolet

0.87

ounty

0.85

obe

0.85

keyes

0.83

bern

0.82

rosse

0.80

irc

0.79

Activations Density 0.030%

names or phrases containing the word "Jac" with varying levels of relevance indicated by activation values

mentions of specific names, particularly "Jac" and "Coch."

No Comments

No Known Activations