INDEX

Explanations

phrases indicating possession or obligation, often in the form of "have to" or "must have"

oai_token-act-pair · gpt-3.5-turbo

expressions of necessity or possession

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

rep

-0.65

 speeds

-0.65

 breakdown

-0.64

res

-0.63

 yards

-0.61

gas

-0.60

 points

-0.60

 seconds

-0.60

 marks

-0.59

 results

-0.59

POSITIVE LOGITS

have

3.55

Have

1.68

has

1.59

want

1.57

were

1.55

 HAVE

1.52

had

1.52

are

1.48

been

1.42

haven

1.41

Activations Density 0.014%

phrases indicating possession or obligation, often in the form of "have to" or "must have"

expressions of necessity or possession

No Comments

No Known Activations