INDEX

Explanations

references to interactions or communications involving the speaker

oai_token-act-pair · gpt-3.5-turbo

references to personal communication or interactions

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

Negative Logits

¿½

-0.68

rawdownloadcloneembedreportprint

-0.63

 Execution

-0.63

ESV

-0.62

profits

-0.62

ģĸ

-0.61

lance

-0.59

 Wikimedia

-0.59

vi

-0.58

cing

-0.57

POSITIVE LOGITS

 symp

0.75

 asking

0.73

 bluff

0.73

 backstage

0.73

 verbally

0.72

 politely

0.72

 afterward

0.71

 privately

0.71

 questioning

0.71

 orally

0.71

Activations Density 0.139%

references to interactions or communications involving the speaker

references to personal communication or interactions

No Known Activations