INDEX

Explanations

instances of attempted actions

oai_token-act-pair · gpt-3.5-turbo

nouns and phrases that indicate attempts, actions, or outcomes

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

 Fury

-0.71

 Apocalypse

-0.67

 Everest

-0.67

 Helsinki

-0.67

 Bark

-0.66

 Sheridan

-0.64

 Rogue

-0.64

ahime

-0.63

 Slaughter

-0.63

 Pixie

-0.61

POSITIVE LOGITS

 including

0.93

 besides

0.92

 imaginable

0.91

 simultaneously

0.85

 redund

0.82

cies

0.78

prises

0.76

tags

0.76

kat

0.75

clerosis

0.74

Activations Density 0.393%

instances of attempted actions

nouns and phrases that indicate attempts, actions, or outcomes

No Comments

No Known Activations