INDEX

Explanations

institutions, specifically focusing on names of universities and research institutes

oai_token-act-pair · gpt-3.5-turbo

references to various academic or research institutes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

soDeliveryDate

-0.98

 conventions

-0.77

 giveaway

-0.71

 emit

-0.69

 evict

-0.69

 corridors

-0.68

 includ

-0.64

chwitz

-0.64

 rake

-0.63

 repe

-0.62

POSITIVE LOGITS

 Technology

1.14

 Integ

1.00

 Architects

0.98

 Applied

0.97

 Contemporary

0.96

 Forensic

0.96

 Electrical

0.93

 Mathemat

0.93

 Fundamental

0.92

 Molecular

0.90

Activations Density 0.050%

institutions, specifically focusing on names of universities and research institutes

references to various academic or research institutes

No Comments

No Known Activations