INDEX

Explanations

occurrences of the word "max" or phrases related to maximum values

oai_token-act-pair · gpt-3.5-turbo

references to maximum values or limits in a technical context

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

borne

-0.75

sembly

-0.71

gered

-0.69

 Seeds

-0.69

 Forth

-0.68

ilitary

-0.66

actionDate

-0.65

mingham

-0.65

 Straw

-0.64

 Newspaper

-0.63

POSITIVE LOGITS

max

3.49

max

3.32

Max

2.30

MAX

1.98

Max

1.96

MAX

1.96

 maximum

1.92

maximum

1.73

 maxim

1.68

 Maximum

1.59

Activations Density 0.016%

occurrences of the word "max" or phrases related to maximum values

references to maximum values or limits in a technical context

No Comments

No Known Activations