INDEX

Explanations

references to public endorsements by individuals or organizations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "endorsed" 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 @ 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

 Caps

-0.67

istic

-0.65

assic

-0.64

 Strait

-0.64

istics

-0.64

Torrent

-0.64

rug

-0.64

ahime

-0.63

IRO

-0.63

 unprotected

-0.62

POSITIVE LOGITS

 endorsing

1.15

 endorse

1.05

 endorsements

1.01

 endorsed

0.96

 endorsement

0.89

eering

0.84

orses

0.79

ablo

0.77

hovah

0.77

clinton

0.74

Activations Density 0.020%

references to public endorsements by individuals or organizations

occurrences of the word "endorsed" and its variants

No Comments

No Known Activations