INDEX

Explanations

actions related to eating

oai_token-act-pair · gpt-3.5-turbo

actions related to eating and social interactions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pects

-0.63

Reviewer

-0.62

 Timeline

-0.61

ocus

-0.58

etheless

-0.56

 RELE

-0.55

NetMessage

-0.55

spective

-0.54

 nutshell

-0.54

 notably

-0.54

POSITIVE LOGITS

 toilet

0.67

 refill

0.64

 snack

0.63

 washing

0.62

 chewing

0.62

 saline

0.60

 toilets

0.60

cig

0.60

 drink

0.60

dye

0.58

Activations Density 2.489%

actions related to eating

actions related to eating and social interactions

No Comments

No Known Activations