INDEX

Explanations

instances of the word "too" used to indicate excess or extremity

oai_token-act-pair · gpt-3.5-turbo

phrases indicating excessiveness or going beyond limits

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

igi

-0.69

inarily

-0.68

idia

-0.67

issance

-0.66

riots

-0.66

yre

-0.64

isode

-0.62

tu

-0.62

quist

-0.61

cu

-0.60

POSITIVE LOGITS

ls

0.85

 much

0.82

 many

0.75

lax

0.74

çĭ

0.73

 overboard

0.73

 risky

0.71

oths

0.70

len

0.69

BuyableInstoreAndOnline

0.68

Activations Density 0.043%

instances of the word "too" used to indicate excess or extremity

phrases indicating excessiveness or going beyond limits

No Comments

No Known Activations