INDEX

Explanations

contractions followed by the word "t"

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Matters

-0.62

side

-0.57

 direction

-0.57

 Signs

-0.56

 Sands

-0.56

 Shrine

-0.56

heads

-0.55

 Torch

-0.55

 Maritime

-0.55

 directions

-0.55

POSITIVE LOGITS

 afford

1.25

weet

0.94

urtle

0.88

ruly

0.87

urtles

0.86

adian

0.85

ogether

0.85

reat

0.85

entimes

0.83

omp

0.80

Activations Density 0.027%

contractions followed by the word "t"

No Comments

No Known Activations