INDEX

Explanations

adjectives related to important and complex topics

oai_token-act-pair · gpt-3.5-turbo

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

trained

-0.80

Cola

-0.77

blers

-0.72

ressor

-0.67

stores

-0.65

mia

-0.63

esters

-0.63

bil

-0.63

VERT

-0.63

ambo

-0.62

POSITIVE LOGITS

 topic

2.00

 topics

1.88

 issue

1.77

 issues

1.73

 matters

1.68

 matter

1.67

 subject

1.64

issues

1.63

 questions

1.54

 question

1.52

Activations Density 0.568%

adjectives related to important and complex topics

No Comments

No Known Activations