INDEX

Explanations

words related to family members or connections

oai_token-act-pair · gpt-3.5-turbo

references to family members and relatives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Cola

-0.80

lite

-0.78

Effective

-0.74

inem

-0.66

yss

-0.66

okin

-0.65

Indust

-0.65

oker

-0.65

atan

-0.65

lins

-0.64

POSITIVE LOGITS

 relatives

1.02

hips

0.96

hetical

0.83

hood

0.81

 reunion

0.80

lia

0.78

ships

0.75

ilial

0.75

 abroad

0.75

 reun

0.74

Activations Density 0.082%

words related to family members or connections

references to family members and relatives

No Comments

No Known Activations