INDEX

Explanations

software and technology-related terms and actions

oai_token-act-pair · gpt-3.5-turbo

references to software features and updates

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

 narratives

-0.77

 Tours

-0.72

 Kissinger

-0.66

 Royals

-0.66

 Salman

-0.65

 Bundy

-0.64

anners

-0.64

 speeches

-0.62

 Oswald

-0.62

orthern

-0.61

POSITIVE LOGITS

 deprecated

1.22

 discontinued

0.99

installed

0.93

 obsolete

0.92

Created

0.88

 ported

0.86

 proprietary

0.86

 plugin

0.86

 usability

0.84

 optimized

0.84

Activations Density 0.858%

software and technology-related terms and actions

references to software features and updates

No Comments

No Known Activations