INDEX

Explanations

phrases related to the legality or intention of actions

oai_token-act-pair · gpt-3.5-turbo

phrases asserting negation or denial

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

raph

-0.69

angs

-0.65

jas

-0.64

zens

-0.64

 swarm

-0.64

 Robots

-0.63

 Falling

-0.62

Surv

-0.62

 resil

-0.62

ogy

-0.61

POSITIVE LOGITS

 discussed

1.17

 communicated

1.17

 done

1.15

 vetted

1.11

 undertaken

1.11

 instituted

1.07

 unintentional

1.06

 contemplated

1.05

 negotiated

1.05

 conveyed

1.05

Activations Density 0.243%

phrases related to the legality or intention of actions

phrases asserting negation or denial

No Comments

No Known Activations