INDEX

Explanations

mentions of 'respond' in a given context, indicating a focus on reactions or lack thereof

oai_token-act-pair · gpt-3.5-turbo

instances of the word "respond" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

cutting

-0.83

metal

-0.71

rome

-0.69

Bav

-0.67

marks

-0.65

our

-0.65

 Damn

-0.64

 Grab

-0.64

Hav

-0.63

 Fant

-0.63

POSITIVE LOGITS

 respond

1.10

reply

0.96

 responders

0.95

 responses

0.94

 responding

0.86

 receptive

0.85

respond

0.84

 responsiveness

0.83

ively

0.83

 thereto

0.82

Activations Density 0.016%

mentions of 'respond' in a given context, indicating a focus on reactions or lack thereof

instances of the word "respond" in various contexts

No Comments

No Known Activations