INDEX

Explanations

references to being officially listed or registered

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "registered" in various contexts

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

wich

-0.84

assi

-0.76

ï¸ı

-0.75

ilitarian

-0.75

fare

-0.68

lihood

-0.65

 Lilly

-0.65

ilt

-0.64

ï¸

-0.64

perty

-0.63

POSITIVE LOGITS

 trademarks

0.93

ifying

0.78

 registered

0.77

emark

0.76

registered

0.75

 irregularities

0.74

 registration

0.70

 trademark

0.70

 paperwork

0.69

essions

0.68

Activations Density 0.013%

references to being officially listed or registered

occurrences of the word "registered" in various contexts

No Comments

No Known Activations