INDEX

Explanations

instances of a specific term, specifically the word "Measure."

oai_token-act-pair · gpt-3.5-turbo

formal terms related to measurement or evaluation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 caloric

-0.70

merce

-0.67

*/(

-0.67

 colle

-0.67

 belong

-0.66

cedented

-0.66

 agre

-0.64

 masse

-0.63

 distingu

-0.62

xual

-0.61

POSITIVE LOGITS

iti

0.72

abase

0.71

ulp

0.69

nexus

0.67

Requirements

0.66

allo

0.65

OEM

0.63

Motor

0.63

tips

0.62

iple

0.62

Activations Density 0.000%

instances of a specific term, specifically the word "Measure."

formal terms related to measurement or evaluation

No Comments

No Known Activations