INDEX

Explanations

specific predictions or statements about the future

oai_token-act-pair · gpt-3.5-turbo

phrases asserting future occurrences or predictions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Doing

-0.63

 preparing

-0.60

 pressuring

-0.60

 projecting

-0.59

 DISTR

-0.58

 stripping

-0.58

 checking

-0.57

 towed

-0.57

 converting

-0.56

 undertaking

-0.56

POSITIVE LOGITS

abouts

0.97

be

0.85

enty

0.85

emonium

0.81

ivably

0.79

 arise

0.77

always

0.77

upon

0.76

arcity

0.73

berra

0.71

Activations Density 0.052%

specific predictions or statements about the future

phrases asserting future occurrences or predictions

No Comments

No Known Activations