INDEX

Explanations

dates followed by blog post times

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Posted" followed by a date or time reference

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

atern

-0.81

ugal

-0.80

ively

-0.73

ifted

-0.72

fell

-0.72

isky

-0.71

ishi

-0.71

ppard

-0.71

roo

-0.70

stood

-0.70

POSITIVE LOGITS

Posted

0.97

 Thumbnails

0.83

 Comments

0.73

monton

0.73

vertis

0.72

 Prediction

0.67

 Helpful

0.66

itors

0.66

 siege

0.65

erick

0.65

Activations Density 0.008%

dates followed by blog post times

instances of the word "Posted" followed by a date or time reference

No Comments

No Known Activations