INDEX

Explanations

occurrences of the word "Full" followed by a number rating

oai_token-act-pair · gpt-3.5-turbo

the term "Full" in various contexts and frequencies

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

åĤ

-0.83

Tup

-0.74

Wars

-0.73

*/(

-0.72

achu

-0.68

ovan

-0.67

å¥

-0.67

ques

-0.65

estern

-0.64

Downloadha

-0.64

POSITIVE LOGITS

erton

0.99

Full

0.82

enhagen

0.82

 fled

0.80

 Full

0.80

 Coverage

0.76

ature

0.74

iaries

0.74

bright

0.73

ģ«

0.71

Activations Density 0.004%

occurrences of the word "Full" followed by a number rating

the term "Full" in various contexts and frequencies

No Comments

No Known Activations

occurrences of the word "Full" followed by a number rating

the term "Full" in various contexts and frequencies

No Comments

No Known Activations