INDEX

Explanations

references to technology products, specifically the Nexus line

oai_token-act-pair · gpt-3.5-turbo

references to Nexus products or concepts

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

ischer

-1.03

lasses

-0.99

lass

-0.95

stroke

-0.81

loss

-0.80

boarding

-0.79

fully

-0.77

netic

-0.76

iage

-0.76

akespe

-0.75

POSITIVE LOGITS

umeric

0.78

mods

0.77

peria

0.74

 Mods

0.73

Mods

0.71

ifier

0.70

ifiers

0.67

 Reborn

0.67

ity

0.66

 Nexus

0.65

Activations Density 0.028%

references to technology products, specifically the Nexus line

references to Nexus products or concepts

No Comments

No Known Activations