INDEX

Explanations

phrases or sentences containing the word "tables" with varying levels of relevance indicated by the activation values

oai_token-act-pair · gpt-3.5-turbo

references to tables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

eworld

-0.70

vernment

-0.69

urt

-0.67

ship

-0.66

trak

-0.65

=~=~

-0.64

PLIED

-0.64

Pacific

-0.64

 Channel

-0.63

chie

-0.63

POSITIVE LOGITS

 tables

1.55

 Tables

1.16

 table

1.09

poons

1.08

table

0.94

cloth

0.87

 shelves

0.84

Table

0.83

 TABLE

0.82

ql

0.81

Activations Density 0.004%

phrases or sentences containing the word "tables" with varying levels of relevance indicated by the activation values

references to tables

No Comments

No Known Activations