INDEX

Explanations

passive constructions where something is happening to an entity or object

oai_token-act-pair · gpt-3.5-turbo

the word "being" in various contexts

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

 srfAttach

-0.67

 Relief

-0.63

 Wick

-0.61

 Tears

-0.60

Wid

-0.60

 Taxes

-0.59

 Mobility

-0.59

 VIDEOS

-0.59

 Explosion

-0.59

 upbringing

-0.57

POSITIVE LOGITS

 able

1.01

 judged

0.88

ijing

0.86

 assessed

0.84

 evaluated

0.84

 eaten

0.83

 treated

0.82

 replaced

0.80

 transported

0.79

ardy

0.79

Activations Density 0.051%

passive constructions where something is happening to an entity or object

the word "being" in various contexts

No Comments

No Known Activations