INDEX

Explanations

symbols or special characters indicating lists

oai_token-act-pair · gpt-3.5-turbo

bullet points or list indicators

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ierre

-0.78

 graz

-0.77

erer

-0.75

othal

-0.75

udic

-0.71

 Tanz

-0.67

nuts

-0.67

enthal

-0.64

 resent

-0.63

bris

-0.63

POSITIVE LOGITS

Â·Â·

1.14

âĢ¢âĢ¢

0.89

Â·

0.85

sim

0.83

¼

0.82

âĢ¢âĢ¢âĢ¢âĢ¢

0.82

NET

0.78

Pg

0.76

 Reason

0.74

¾

0.74

Activations Density 0.005%

symbols or special characters indicating lists

bullet points or list indicators

No Comments

No Known Activations