INDEX

Explanations

questions starting with "Should" or "Could"

oai_token-act-pair · gpt-3.5-turbo

negations or questions regarding choices and rights

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Photographer

-0.68

 halves

-0.64

 Craigslist

-0.61

EStreamFrame

-0.61

 Toad

-0.60

Dialog

-0.59

 dracon

-0.58

 winding

-0.57

 wilderness

-0.57

 Ãľ

-0.56

POSITIVE LOGITS

aught

1.03

ovation

1.00

astics

0.94

ails

0.94

onsense

0.94

aturally

0.91

ocent

0.90

ough

0.89

itely

0.88

asty

0.88

Activations Density 0.021%

questions starting with "Should" or "Could"

negations or questions regarding choices and rights

No Comments

No Known Activations