INDEX

Explanations

syntax patterns typically used in coding or technical contexts

oai_token-act-pair · gpt-3.5-turbo

delimiters and structural elements in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ensis

-0.90

 toget

-0.75

hem

-0.74

aciously

-0.68

 ammon

-0.66

 chloride

-0.65

 thor

-0.65

gou

-0.65

 leasing

-0.64

lease

-0.64

POSITIVE LOGITS

 Majesty

0.90

 Commons

0.89

Tube

0.88

 Subtle

0.88

 Decay

0.87

Va

0.87

Zero

0.85

 Honour

0.84

 Prevention

0.84

Sex

0.82

Activations Density 0.684%

syntax patterns typically used in coding or technical contexts

delimiters and structural elements in a text

No Comments

No Known Activations