INDEX

Explanations

names or terms related to politics and individuals

oai_token-act-pair · gpt-3.5-turbo

names of people or notable individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ModLoader

-0.83

theless

-0.61

å§«

-0.59

 Galileo

-0.58

âĶĢâĶĢ

-0.57

ãĥŁ

-0.57

etheless

-0.56

ashtra

-0.56

 Fancy

-0.56

 Eleanor

-0.55

POSITIVE LOGITS

antz

0.71

burn

0.71

linger

0.71

zen

0.70

lett

0.69

beck

0.68

zer

0.67

berg

0.66

erman

0.65

mann

0.65

Activations Density 0.281%

names or terms related to politics and individuals

names of people or notable individuals

No Comments

No Known Activations