INDEX

Explanations

sequences indicating a repeated action or event

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "consecutive" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

rets

-0.72

forts

-0.69

Pros

-0.67

craft

-0.65

UX

-0.63

oras

-0.62

ffer

-0.61

ourcing

-0.61

ree

-0.60

stuff

-0.59

POSITIVE LOGITS

 consecutive

3.50

 successive

2.54

 consecut

2.10

 straight

1.43

 uninterrupted

1.40

straight

1.38

 simultaneous

1.36

 sequential

1.26

secut

1.23

 cumulative

1.21

Activations Density 0.021%

sequences indicating a repeated action or event

occurrences of the word "consecutive" and related terms

No Comments

No Known Activations