INDEX

Explanations

words related to various approaches or methods

oai_token-act-pair · gpt-3.5-turbo

different strategies or methods used in various 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

gin

-0.73

grave

-0.70

ãĥ©ãĥ³

-0.68

arus

-0.64

bered

-0.64

reported

-0.63

 nurs

-0.62

Wak

-0.62

rake

-0.61

cakes

-0.61

POSITIVE LOGITS

 approach

0.88

ologies

0.88

 toward

0.88

 Approach

0.83

ahime

0.82

 towards

0.81

rait

0.77

 thereto

0.77

ivist

0.76

able

0.75

Activations Density 0.027%

words related to various approaches or methods

different strategies or methods used in various contexts

No Comments

No Known Activations