INDEX

Explanations

phrases related to official statements or declarations

oai_token-act-pair · gpt-3.5-turbo

statements involving official remarks and responses from individuals in authority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

TPPStreamerBot

-0.80

 underest

-0.76

 hindsight

-0.74

 underestimated

-0.74

iencies

-0.72

ĵĺ

-0.70

 scept

-0.69

 hampered

-0.66

dding

-0.65

 doubted

-0.65

POSITIVE LOGITS

 wrote

1.38

 write

1.10

 reads

1.09

 statement

1.09

 writes

1.05

wrote

1.04

 read

1.02

 letter

0.99

letter

0.99

 written

0.99

Activations Density 0.230%

phrases related to official statements or declarations

statements involving official remarks and responses from individuals in authority

No Comments

No Known Activations