INDEX

Explanations

expressions of optimism or confidence

oai_token-act-pair · gpt-3.5-turbo

expressions of certainty or predictions about future outcomes

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

said

-0.71

DAQ

-0.71

RD

-0.66

told

-0.65

VPN

-0.61

OAD

-0.60

 Sched

-0.60

inia

-0.58

ENDED

-0.58

POSITIVE LOGITS

 someday

1.30

 eventually

1.23

 prevail

1.20

 succeed

1.18

be

1.15

 soon

1.12

 somehow

1.12

 suffice

1.06

never

1.03

 never

1.02

Activations Density 0.184%

expressions of optimism or confidence

expressions of certainty or predictions about future outcomes

No Comments

No Known Activations