INDEX

Explanations

words relating to health risks and potential harm

oai_token-act-pair · gpt-3.5-turbo

terms related to the negative effects of substances or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

rollers

-0.69

ciples

-0.66

ahs

-0.65

zar

-0.63

doms

-0.62

ilde

-0.62

gres

-0.62

IDA

-0.61

quart

-0.60

elle

-0.59

POSITIVE LOGITS

 insofar

0.95

 enough

0.93

 owing

0.87

 unless

0.86

 compared

0.85

 towards

0.80

 because

0.79

 deterrent

0.76

 toward

0.75

 against

0.74

Activations Density 0.245%

words relating to health risks and potential harm

terms related to the negative effects of substances or actions

No Comments

No Known Activations