INDEX

Explanations

text phrases that have a specific symbol sequence (âĢ)

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "titled."

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

 range

-0.79

 detached

-0.71

 sear

-0.71

 shroud

-0.71

 bombard

-0.69

creen

-0.69

 infring

-0.67

 shelf

-0.65

 multiplying

-0.65

 division

-0.64

POSITIVE LOGITS

ª

1.42

ı

1.20

ł

1.20

¹

1.18

Ĵ

1.14

«

1.10

Ķ

1.09

¡

1.09

ĳ

1.09

§

1.04

Activations Density 0.129%

text phrases that have a specific symbol sequence (âĢ)

occurrences of the word "titled."

No Comments

No Known Activations