INDEX

Explanations

phrases related to potential emergencies or significant events that may require preparation or response

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditional scenarios or events related to emergencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

kefeller

-0.79

roxy

-0.73

roots

-0.72

minus

-0.70

hua

-0.70

mostly

-0.70

ilib

-0.68

cigarettes

-0.67

pha

-0.67

pees

-0.67

POSITIVE LOGITS

 unforeseen

1.18

 malf

1.17

 malfunction

1.17

 emergencies

1.15

 misfortune

1.13

 breach

1.07

 unexpectedly

1.07

 disagreement

1.06

 conflict

1.06

 lapse

1.03

Activations Density 0.533%

phrases related to potential emergencies or significant events that may require preparation or response

phrases indicating conditional scenarios or events related to emergencies

No Comments

No Known Activations