INDEX

Explanations

names of individuals named Kenneth

oai_token-act-pair · gpt-3.5-turbo

references to individuals named Kenneth

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vous

-0.85

odcast

-0.73

drive

-0.72

cells

-0.70

å¥³

-0.68

resso

-0.67

mith

-0.66

Ö¼

-0.65

etition

-0.65

expression

-0.65

POSITIVE LOGITS

 Kenneth

0.92

 Starr

0.89

owan

0.82

 Anger

0.80

 Fernandez

0.77

 Bran

0.77

 Cole

0.74

 Thomson

0.72

 Thompson

0.71

 Choi

0.71

Activations Density 0.014%

names of individuals named Kenneth

references to individuals named Kenneth

No Comments

No Known Activations