INDEX

Explanations

mentions of fatal events or incidents

oai_token-act-pair · gpt-3.5-turbo

references to fatal incidents or deaths

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

orthy

-0.80

here

-0.73

annis

-0.73

arta

-0.71

arte

-0.69

arity

-0.69

atters

-0.69

adr

-0.68

enda

-0.68

 Remastered

-0.67

POSITIVE LOGITS

istically

1.13

istic

1.09

ities

1.02

 fatal

0.99

 flaw

0.98

ized

0.97

ist

0.95

ism

0.91

izes

0.89

 overdose

0.86

Activations Density 0.016%

mentions of fatal events or incidents

references to fatal incidents or deaths

No Comments

No Known Activations