INDEX

Explanations

phrases related to legal proceedings

oai_token-act-pair · gpt-3.5-turbo

repetitive conjunctions and connectors in sentences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

bub

-0.60

Base

-0.59

gest

-0.55

burgh

-0.54

ãĥ¼ãĥ

-0.53

 Seym

-0.52

zu

-0.52

Class

-0.52

abul

-0.52

kel

-0.51

POSITIVE LOGITS

 secondly

1.04

 therefore

0.76

 consequently

0.75

 thence

0.75

 dont

0.71

rogens

0.71

 thats

0.71

 didnt

0.70

 thus

0.70

 hence

0.69

Activations Density 0.357%

phrases related to legal proceedings

repetitive conjunctions and connectors in sentences

No Comments

No Known Activations