INDEX

Explanations

phrases related to criticism or negative attention towards a person or entity

oai_token-act-pair · gpt-3.5-turbo

phrases related to criticism or controversy surrounding individuals or entities

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

rave

-0.71

 Nope

-0.66

olute

-0.63

mu

-0.62

 Norn

-0.60

hibition

-0.59

 cures

-0.58

 vantage

-0.57

Yep

-0.57

eworks

-0.55

POSITIVE LOGITS

 lately

0.90

 recently

0.82

for

0.81

 stemming

0.78

 domestically

0.77

 nationally

0.74

 internationally

0.73

 levied

0.72

due

0.72

 owing

0.72

Activations Density 0.168%

phrases related to criticism or negative attention towards a person or entity

phrases related to criticism or controversy surrounding individuals or entities

No Comments

No Known Activations