INDEX

Explanations

technical terms or commands related to software or technology

oai_token-act-pair · gpt-3.5-turbo

terms related to technology, organizations, and structural classifications

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

 Exploration

-0.61

 Helsinki

-0.61

vich

-0.59

 Incre

-0.57

 Democracy

-0.57

 Mayo

-0.55

grim

-0.55

00200000

-0.55

azard

-0.55

politics

-0.54

POSITIVE LOGITS

 imaginable

1.05

hips

1.01

mith

0.85

 besides

0.82

paces

0.81

hops

0.77

 consecut

0.75

 vying

0.75

 including

0.71

 redund

0.70

Activations Density 0.632%

technical terms or commands related to software or technology

terms related to technology, organizations, and structural classifications

No Comments

No Known Activations