INDEX

Explanations

metrical or measurement-related terms, especially focusing on attributes such as quality, characteristics, or standards

oai_token-act-pair · gpt-3.5-turbo

terms related to measurement or quality assessment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hus

-0.82

overs

-0.70

ey

-0.70

KEN

-0.69

idem

-0.69

sle

-0.68

 Hyde

-0.68

folk

-0.67

egg

-0.67

gotten

-0.67

POSITIVE LOGITS

 caliber

1.10

 calib

1.00

uracy

0.83

 inacc

0.79

illon

0.78

anguage

0.77

 ammunition

0.76

caliber

0.75

 specimen

0.75

orsi

0.75

Activations Density 0.009%

metrical or measurement-related terms, especially focusing on attributes such as quality, characteristics, or standards

terms related to measurement or quality assessment

No Comments

No Known Activations