INDEX

Explanations

statements containing the word "not"

oai_token-act-pair · gpt-3.5-turbo

negations or instances of non-authorization in statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

opia

-0.65

 aspiration

-0.62

 Puzzles

-0.62

 supremacy

-0.62

soDeliveryDate

-0.62

 generational

-0.60

 nightmare

-0.59

 Solitaire

-0.59

 Conversation

-0.59

Pop

-0.58

POSITIVE LOGITS

 disclosed

1.13

 immediately

1.11

 disclose

1.02

 disclosing

0.99

 divul

0.96

 formally

0.95

ifies

0.93

 explicitly

0.89

 publicly

0.89

 discl

0.89

Activations Density 0.120%

statements containing the word "not"

negations or instances of non-authorization in statements

No Comments

No Known Activations