INDEX

Explanations

instances of the word "Four"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Four"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

rad

-0.66

utm

-0.62

dmg

-0.61

 spam

-0.60

ILA

-0.59

etting

-0.58

 confuse

-0.58

erv

-0.57

 answ

-0.56

err

-0.56

POSITIVE LOGITS

 Four

3.45

Four

2.36

Six

2.28

 Eight

2.19

 Five

2.16

 Three

2.14

Fif

2.09

 Seven

1.89

 Forty

1.84

 Twelve

1.82

Activations Density 0.008%

instances of the word "Four"

references to the term "Four"

No Comments

No Known Activations