INDEX

Explanations

proper nouns, potentially names of people, in a list

oai_token-act-pair · gpt-3.5-turbo

key names of people 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

tenance

-0.87

emort

-0.79

iquid

-0.76

inarily

-0.76

 citiz

-0.76

ometimes

-0.74

 spectrum

-0.73

WER

-0.72

onym

-0.71

 subjective

-0.71

POSITIVE LOGITS

ken

0.74

pton

0.74

 Juda

0.72

Jr

0.68

 Bagg

0.68

Cul

0.67

 Architects

0.66

 Casting

0.65

 joins

0.65

Es

0.64

Activations Density 3.090%

proper nouns, potentially names of people, in a list

key names of people in various contexts

No Comments

No Known Activations