INDEX

Explanations

phrases related to actions or events being attributed or associated with someone

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "has" in different contexts

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

iscover

-0.67

ercise

-0.59

iling

-0.57

 Pair

-0.57

 recall

-0.56

 urge

-0.56

 Discuss

-0.55

du

-0.55

 Guess

-0.54

 solicitation

-0.54

POSITIVE LOGITS

 happened

1.11

 been

1.08

 nothing

1.05

 arisen

1.03

nothing

1.00

 occurred

0.97

 ramifications

0.96

 implications

0.94

 transpired

0.92

bara

0.90

Activations Density 0.139%

phrases related to actions or events being attributed or associated with someone

instances of the verb "has" in different contexts

No Comments

No Known Activations