INDEX

Explanations

phrases related to motivating or inciting actions or behaviors

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate motivation or inspiring actions towards a goal

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

ary

-0.75

listed

-0.65

ended

-0.65

 briefed

-0.63

fee

-0.62

arity

-0.62

 fucked

-0.60

awan

-0.60

handled

-0.59

oken

-0.59

POSITIVE LOGITS

 undertake

1.23

 pursue

1.22

 rethink

1.18

 seek

1.18

 innovate

1.05

 abandon

1.03

 explore

1.03

 migrate

1.02

 contemplate

1.02

 engage

1.02

Activations Density 0.199%

phrases related to motivating or inciting actions or behaviors

phrases that indicate motivation or inspiring actions towards a goal

No Comments

No Known Activations