INDEX

Explanations

mentions of small amounts or degrees of something

oai_token-act-pair · gpt-3.5-turbo

phrases mentioning a "little bit" or slight amounts

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

ħĭ

-1.00

¥µ

-0.77

aring

-0.72

Ń·

-0.69

rights

-0.67

andra

-0.67

ãĤ¤ãĥĪ

-0.66

 Rath

-0.66

åĤ

-0.64

furt

-0.64

POSITIVE LOGITS

 aspirin

0.75

(<

0.75

 overboard

0.74

 forgiven

0.70

terness

0.65

emark

0.65

lycer

0.64

 immersed

0.63

plug

0.62

bang

0.61

Activations Density 0.032%

mentions of small amounts or degrees of something

phrases mentioning a "little bit" or slight amounts

No Comments

No Known Activations