INDEX

Explanations

words related to threats or actions that may cause harm or danger

oai_token-act-pair · gpt-3.5-turbo

words related to threats or danger

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

mys

-0.94

gart

-0.92

mot

-0.80

uesday

-0.77

puted

-0.76

angles

-0.75

area

-0.74

éĸ

-0.73

arist

-0.72

aho

-0.72

POSITIVE LOGITS

 adolesc

0.77

 threats

0.76

 challeng

0.76

 endanger

0.75

 retaliation

0.72

cffff

0.71

 threatened

0.70

Ã±o

0.69

lessly

0.68

lihood

0.68

Activations Density 0.023%

words related to threats or actions that may cause harm or danger

words related to threats or danger

No Comments

No Known Activations