INDEX

Explanations

terms related to significant discoveries or surprising information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "revelation" and its variants

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

asus

-0.87

animate

-0.78

osh

-0.77

ope

-0.77

veyard

-0.75

ename

-0.74

erate

-0.73

cul

-0.72

keye

-0.72

clock

-0.67

POSITIVE LOGITS

 revelations

1.19

 revelation

1.12

 Revelations

0.87

Leaks

0.84

 reveals

0.81

 reveal

0.80

 disclosures

0.78

 revealed

0.77

 concerning

0.76

 revealing

0.76

Activations Density 0.026%

terms related to significant discoveries or surprising information

instances of the word "revelation" and its variants

No Comments

No Known Activations