INDEX

Explanations

scientific or technical terms related to mechanisms or processes

oai_token-act-pair · gpt-3.5-turbo

references to mechanisms or processes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hiro

-0.72

 Parables

-0.72

minster

-0.70

iaries

-0.70

atana

-0.67

uden

-0.66

oning

-0.65

ield

-0.64

 Childhood

-0.64

 Cardinals

-0.64

POSITIVE LOGITS

 mechanism

0.94

 Mechan

0.90

 mechanisms

0.88

utics

0.86

witz

0.85

 whereby

0.85

ptoms

0.84

 workings

0.77

 symmetry

0.73

ulic

0.71

Activations Density 0.023%

scientific or technical terms related to mechanisms or processes

references to mechanisms or processes

No Comments

No Known Activations