INDEX

Explanations

numbers or data points within sentences

oai_token-act-pair · gpt-3.5-turbo

recommendations for reading material or content to explore

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 subcontract

-0.75

andowski

-0.72

antle

-0.67

 quotas

-0.66

 lobb

-0.65

Alc

-0.62

 announces

-0.61

 euth

-0.60

 bills

-0.60

 masc

-0.60

POSITIVE LOGITS

 reading

1.20

Reading

1.16

 read

1.11

reader

1.08

 CLICK

1.03

 READ

1.02

 download

1.00

 viewing

0.99

read

0.99

download

0.98

Activations Density 0.383%

numbers or data points within sentences

recommendations for reading material or content to explore

No Comments

No Known Activations