INDEX

Explanations

descriptions or discussions about visual figures or sculptures

oai_token-act-pair · gpt-3.5-turbo

references to figurines or figures

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

bey

-0.80

é¾

-0.79

velength

-0.75

cyclopedia

-0.72

reath

-0.72

ruciating

-0.71

bnb

-0.71

ory

-0.70

termination

-0.70

umenthal

-0.69

POSITIVE LOGITS

 skating

0.95

 prominently

0.87

head

0.84

figure

0.81

 figure

0.77

enance

0.75

heads

0.74

maker

0.72

 macros

0.69

 sculpt

0.69

Activations Density 0.021%

descriptions or discussions about visual figures or sculptures

references to figurines or figures

No Comments

No Known Activations