INDEX

Explanations

skills and traits that are useful or desirable for various purposes

oai_token-act-pair · gpt-3.5-turbo

terms related to skills and positive attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

oulos

-0.77

 sidx

-0.75

 Lumpur

-0.73

usalem

-0.71

thumbnails

-0.69

doms

-0.69

racuse

-0.68

asers

-0.67

azeera

-0.66

assies

-0.66

POSITIVE LOGITS

 requiring

0.83

 used

0.77

 prized

0.76

 invented

0.76

 reliever

0.75

 utilized

0.75

 nowadays

0.74

 whereas

0.71

 favored

0.69

 unto

0.69

Activations Density 0.303%

skills and traits that are useful or desirable for various purposes

terms related to skills and positive attributes

No Comments

No Known Activations