INDEX

Explanations

words related to universities, institutions, and events

oai_token-act-pair · gpt-3.5-turbo

names of institutions and publications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 jokes

-0.85

 begg

-0.80

 background

-0.78

 reapp

-0.77

jriwal

-0.77

metadata

-0.76

dyl

-0.74

ebin

-0.73

uphem

-0.72

 selfies

-0.70

POSITIVE LOGITS

University

1.02

 Institute

0.99

 University

0.99

Kn

0.96

Tec

0.93

 Sports

0.92

 Corporation

0.92

 College

0.89

 Precision

0.89

 Publishing

0.88

Activations Density 0.835%

words related to universities, institutions, and events

names of institutions and publications

No Comments

No Known Activations