INDEX

Explanations

instances where the concept of returning or resuming a previous state or action is mentioned

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a return or going back to a previous state or situation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 omission

-0.67

 heartbeat

-0.61

 Liang

-0.58

PLA

-0.57

Haf

-0.57

 giveaway

-0.56

 Belg

-0.55

 bene

-0.54

 banks

-0.54

 prominently

-0.53

POSITIVE LOGITS

 sidx

0.79

anship

0.75

cloneembedreportprint

0.74

/+

0.74

Ïī

0.73

*=-

0.73

involved

0.72

ocobo

0.72

actionDate

0.72

 traction

0.71

Activations Density 0.084%

instances where the concept of returning or resuming a previous state or action is mentioned

phrases that indicate a return or going back to a previous state or situation

No Comments

No Known Activations