INDEX

Explanations

phrases expressing personal preferences or recommendations

oai_token-act-pair · gpt-3.5-turbo

expressions of desire or intention

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

 Gutenberg

-0.74

Kag

-0.69

çİĭ

-0.60

 Basics

-0.60

 Anim

-0.59

 Kinder

-0.59

Cook

-0.59

 PLUS

-0.58

DATA

-0.57

 Material

-0.56

POSITIVE LOGITS

 gladly

1.04

 advise

0.91

 dearly

0.89

 definitely

0.88

 never

0.86

 recommend

0.86

 sugg

0.85

 assume

0.83

 summarize

0.83

 certainly

0.83

Activations Density 0.083%

phrases expressing personal preferences or recommendations

expressions of desire or intention

No Comments

No Known Activations