INDEX

Explanations

comparative statements indicating superiority or excellence

oai_token-act-pair · gpt-3.5-turbo

terms related to surpassing or exceeding benchmarks or limits

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

loc

-0.84

llo

-0.72

ll

-0.72

arty

-0.71

 Fein

-0.67

mun

-0.63

hair

-0.63

random

-0.63

xious

-0.61

 orientation

-0.60

POSITIVE LOGITS

 surpassed

3.25

 surpass

3.18

 eclips

2.45

 exceeded

2.03

 exceed

1.76

 exceeds

1.71

 eclipse

1.70

 overtake

1.59

 outper

1.58

 exceeding

1.53

Activations Density 0.031%

comparative statements indicating superiority or excellence

terms related to surpassing or exceeding benchmarks or limits

No Comments

No Known Activations