INDEX

Explanations

family relationships

oai_token-act-pair · gpt-3.5-turbo

references to in-laws and family relations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 investigators

-0.66

 itch

-0.66

ħĭ

-0.65

 convol

-0.65

edIn

-0.63

ramid

-0.61

 athlet

-0.61

 enthusi

-0.60

 authorities

-0.60

 smugglers

-0.60

POSITIVE LOGITS

chief

1.00

progress

0.98

sight

0.94

terms

0.93

depth

0.92

arms

0.88

speech

0.86

confidence

0.86

dist

0.84

training

0.84

Activations Density 0.036%

family relationships

references to in-laws and family relations

No Comments

No Known Activations