INDEX

Explanations

phrases indicating a response or reaction to something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "respond" and its variations

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

Not in Any Lists

No Comments

Negative Logits

 ILCS

-0.75

utical

-0.72

hered

-0.70

Tool

-0.70

rek

-0.67

sky

-0.66

rette

-0.66

caps

-0.65

lim

-0.65

tool

-0.65

POSITIVE LOGITS

 criticisms

0.97

 criticism

0.97

 requests

0.91

 queries

0.82

 inquiries

0.80

 stimuli

0.78

 allegations

0.78

 critics

0.76

 questions

0.75

 complaints

0.74

Activations Density 0.092%

phrases indicating a response or reaction to something

instances of the word "respond" and its variations

No Comments

No Known Activations