INDEX

Explanations

numeric values

oai_token-act-pair · gpt-3.5-turbo

mathematical or programming assignments and equations

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

 breed

-0.73

 Lauder

-0.73

 resume

-0.69

 casting

-0.69

 limp

-0.66

 alumni

-0.65

 absentee

-0.65

 contrace

-0.65

 coat

-0.64

 braces

-0.64

POSITIVE LOGITS

ãĥīãĥ©ãĤ´ãĥ³

1.09

true

0.98

ãĥ´ãĤ¡

0.96

ãĤ¨ãĥ«

0.94

mc

0.93

CVE

0.93

PsyNetMessage

0.93

false

0.92

lambda

0.92

saf

0.91

Activations Density 0.016%

numeric values

mathematical or programming assignments and equations

No Comments

No Known Activations