INDEX

Explanations

the concept of counting or quantifying

oai_token-act-pair · gpt-3.5-turbo

repetitions of the phrase "a number" followed by a numerical representation of quantity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Despair

-0.68

anium

-0.68

rador

-0.66

inka

-0.65

sein

-0.63

rament

-0.61

agate

-0.61

culosis

-0.61

ovie

-0.61

 Doodle

-0.60

POSITIVE LOGITS

of

0.73

ãĥ¼ãĥ³

0.70

imilar

0.67

encies

0.66

ãĥĥãĤ¯

0.65

otom

0.65

ttes

0.63

xes

0.63

 engagements

0.62

ones

0.61

Activations Density 0.030%

the concept of counting or quantifying

repetitions of the phrase "a number" followed by a numerical representation of quantity

No Comments

No Known Activations