INDEX

Explanations

the word "to" followed by a verb

oai_token-act-pair · gpt-3.5-turbo

phrases that express intention or future actions

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

CTV

-0.82

Reporting

-0.79

Pres

-0.73

cart

-0.72

mere

-0.69

cius

-0.68

checking

-0.67

amps

-0.67

VIDEO

-0.65

Previous

-0.65

POSITIVE LOGITS

be

1.06

 make

0.92

 lose

0.92

 stick

0.91

eat

0.90

try

0.89

get

0.89

 give

0.89

 regress

0.89

 continue

0.88

Activations Density 0.091%

the word "to" followed by a verb

phrases that express intention or future actions

No Comments

No Known Activations