INDEX

Explanations

phrases related to quotes or statements made by various individuals or entities

oai_token-act-pair · gpt-3.5-turbo

mentions of news organizations or press sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 diaper

-0.60

 veter

-0.56

atible

-0.54

 doesnt

-0.54

viz

-0.53

 favourites

-0.52

animate

-0.51

 precon

-0.51

 harms

-0.50

 unfocusedRange

-0.49

POSITIVE LOGITS

0.93

.</

0.81

."

0.77

.).

0.75

 sarcast

0.73

."

0.70

 quoted

0.70

lied

0.69

.]

0.69

 rhet

0.69

Activations Density 0.196%

phrases related to quotes or statements made by various individuals or entities

mentions of news organizations or press sources

No Comments

No Known Activations