INDEX

Explanations

prepositions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "down," indicating decreases or reductions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

-0.65

tein

-0.65

://

-0.64

#$

-0.63

ature

-0.63

âģ

-0.63

ĸļ

-0.61

 Huck

-0.60

cius

-0.60

inka

-0.59

POSITIVE LOGITS

stairs

1.26

graded

1.14

grading

1.05

grades

0.94

river

0.88

pour

0.86

hill

0.85

 stairs

0.85

LOAD

0.85

fall

0.81

Activations Density 0.056%

prepositions

instances of the word "down," indicating decreases or reductions

No Comments

No Known Activations