INDEX

Explanations

references to historical events and dates

oai_token-act-pair · gpt-3.5-turbo

time references and historical events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 emoji

-0.86

 Snapchat

-0.83

 vape

-0.76

(@

-0.75

 ðŁ

-0.71

 ðŁĺ

-0.70

GIF

-0.69

ï¸ı

-0.69

 GoPro

-0.69

 veggies

-0.68

POSITIVE LOGITS

 postwar

0.97

 Yugoslavia

0.92

0.88

 Yugoslav

0.87

0.84

0.83

 Soviet

0.83

Activations Density 2.302%

references to historical events and dates

time references and historical events

No Comments

No Known Activations