INDEX

Explanations

references to the source where an article was originally published

oai_token-act-pair · gpt-3.5-turbo

citations and references to the publication history of articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

respective

-0.70

 thence

-0.69

 preferably

-0.64

 successive

-0.62

gio

-0.62

 hurry

-0.61

rouse

-0.61

atever

-0.59

 pleasing

-0.59

 disg

-0.58

POSITIVE LOGITS

 August

0.83

 July

0.83

 October

0.81

 March

0.81

 November

0.79

 April

0.77

 February

0.76

 September

0.76

 December

0.76

 January

0.76

Activations Density 0.071%

references to the source where an article was originally published

citations and references to the publication history of articles

No Comments

No Known Activations