INDEX

Explanations

references to technology and specific product names, potentially related to processors and devices

oai_token-act-pair · gpt-3.5-turbo

nouns or brand names associated with technology and entertainment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

comings

-0.71

 loose

-0.68

conserv

-0.67

natureconservancy

-0.65

 è£ıè

-0.63

wat

-0.63

angular

-0.63

 overhead

-0.63

htaking

-0.62

anova

-0.60

POSITIVE LOGITS

phia

0.81

ploy

0.80

 coli

0.78

isse

0.75

 Fortune

0.74

ritch

0.70

leanor

0.70

enment

0.68

gypt

0.68

ionage

0.67

Activations Density 0.638%

references to technology and specific product names, potentially related to processors and devices

nouns or brand names associated with technology and entertainment

No Comments

No Known Activations