INDEX

Explanations

terms related to the highest point or peak of something

oai_token-act-pair · gpt-3.5-turbo

references to peaks or high points, both in physical and metaphorical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

angu

-0.79

vor

-0.77

book

-0.74

ellow

-0.72

aret

-0.72

hi

-0.71

ish

-0.71

aved

-0.70

folk

-0.69

mented

-0.69

POSITIVE LOGITS

 apex

0.96

 Apex

0.95

 predator

0.90

xon

0.86

 pinnacle

0.80

reme

0.78

 predators

0.76

iary

0.73

 scorer

0.73

assador

0.72

Activations Density 0.004%

terms related to the highest point or peak of something

references to peaks or high points, both in physical and metaphorical contexts

No Comments

No Known Activations