INDEX

Explanations

technical terms denoting advanced quality or classification

oai_token-act-pair · gpt-3.5-turbo

terms related to grading systems and historical eras

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

IFrame

Link

Negative Logits

pload

-0.85

liga

-0.82

osponsors

-0.81

alus

-0.75

arth

-0.74

pac

-0.71

Shares

-0.71

ideo

-0.70

idas

-0.69

avers

-0.68

POSITIVE LOGITS

quartered

0.76

ipment

0.72

 vessels

0.71

 hybrid

0.71

riors

0.70

 highs

0.70

 hybrids

0.67

 equivalents

0.67

lihood

0.67

 versions

0.66

Activations Density 0.123%

technical terms denoting advanced quality or classification

terms related to grading systems and historical eras

No Known Activations