INDEX

Explanations

references to specific individuals and their interactions

oai_token-act-pair · gpt-3.5-turbo

statements and discussions about events or incidents

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

IFrame

Link

Negative Logits

 accounted

-0.64

ogether

-0.61

 ensures

-0.61

quartered

-0.59

 preserves

-0.58

 depends

-0.57

 depend

-0.56

 stabilized

-0.56

immune

-0.56

 reliant

-0.55

POSITIVE LOGITS

rompt

0.64

 interviewer

0.59

 keynote

0.57

QUEST

0.56

SourceFile

0.56

 famous

0.55

AMA

0.55

 sarcast

0.55

Ask

0.55

Tweet

0.55

Activations Density 1.335%

references to specific individuals and their interactions

statements and discussions about events or incidents

No Known Activations