INDEX

Explanations

language indicating high likelihood or certainty

oai_token-act-pair · gpt-3.5-turbo

the word "almost" and related phrases indicating near certainty or approximation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

nesota

-0.81

ourses

-0.77

ä¸Ń

-0.71

oris

-0.70

nation

-0.69

oland

-0.68

 Dynamics

-0.68

ACTED

-0.66

opsis

-0.66

 è£ıè¦ļéĨĴ

-0.65

POSITIVE LOGITS

 impossible

1.01

 identical

0.97

 indistinguishable

0.94

 certainly

0.88

 universally

0.88

 always

0.84

 nonexistent

0.84

 extinct

0.83

 unrecogn

0.82

 guaranteed

0.80

Activations Density 0.049%

language indicating high likelihood or certainty

the word "almost" and related phrases indicating near certainty or approximation

No Comments

No Known Activations