INDEX

Explanations

words related to actions or processes

oai_token-act-pair · gpt-3.5-turbo

references to "basing" 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 @ 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

 Delicious

-0.78

VICE

-0.73

 Eleven

-0.72

 Quan

-0.71

pelling

-0.70

oice

-0.69

ESSION

-0.69

 Clover

-0.68

NX

-0.68

speak

-0.66

POSITIVE LOGITS

bas

3.91

bas

1.62

Bas

1.50

 bases

1.42

Bas

1.39

 basin

1.36

 baseman

1.08

 base

1.03

 basement

1.00

ber

1.00

Activations Density 0.009%

words related to actions or processes

references to "basing" in various contexts

No Comments

No Known Activations