INDEX

Explanations

words related to computer programming or specific codes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "code" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

esville

-0.84

swer

-0.82

kus

-0.79

athi

-0.78

iatus

-0.78

lication

-0.75

riks

-0.73

 Ples

-0.72

 experien

-0.72

itor

-0.70

POSITIVE LOGITS

 snippet

0.94

velop

0.90

 codes

0.82

 snippets

0.80

ãĥĥãĥī

0.80

codes

0.77

otle

0.77

breakers

0.73

 coded

0.72

breaker

0.72

Activations Density 0.030%

words related to computer programming or specific codes

instances of the word "code" and its variations in different contexts

No Comments

No Known Activations