INDEX

Explanations

words related to documents or written content, particularly focusing on short forms like short films, briefs, and singles

oai_token-act-pair · gpt-3.5-turbo

terms related to short forms or brevity

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

bows

-0.67

 Covenant

-0.61

 Bots

-0.61

 disparate

-0.59

 Peoples

-0.58

jured

-0.56

 earthly

-0.56

 abst

-0.55

 settlements

-0.55

âĹ¼

-0.55

POSITIVE LOGITS

agonist

0.79

otype

0.76

cade

0.75

milo

0.74

izer

0.73

isi

0.73

iser

0.73

izen

0.72

ieth

0.70

 featuring

0.70

Activations Density 0.293%

words related to documents or written content, particularly focusing on short forms like short films, briefs, and singles

terms related to short forms or brevity

No Comments

No Known Activations