INDEX

Explanations

the word "it" followed by a verb indicating an action

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "do it."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

inventoryQuantity

-0.79

Tru

-0.70

ONSORED

-0.66

EStreamFrame

-0.65

being

-0.65

 Sett

-0.61

Whe

-0.61

 Arri

-0.61

ãĤ¦ãĤ¹

-0.61

 Flavoring

-0.61

POSITIVE LOGITS

pez

0.85

alias

0.78

 justice

0.77

job

0.71

 imitate

0.68

 destro

0.67

 legally

0.67

 offline

0.66

alia

0.66

 experiment

0.65

Activations Density 0.084%

the word "it" followed by a verb indicating an action

instances of the phrase "do it."

No Comments

No Known Activations