INDEX

Explanations

instances of the word "proved."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "proved."

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

isha

-0.64

 solicit

-0.63

download

-0.63

robe

-0.59

 pause

-0.55

 Contact

-0.54

contact

-0.54

 Dispatch

-0.54

WM

-0.54

dos

-0.53

POSITIVE LOGITS

 proved

3.31

 proven

2.19

 proves

2.16

 prove

2.07

 proving

1.69

 demonstrated

1.68

proven

1.58

 showed

1.39

 ensured

1.34

 succeeded

1.31

Activations Density 0.007%

instances of the word "proved."

occurrences of the word "proved."

No Comments

No Known Activations