INDEX

Explanations

words related to the desirability of objects, conditions, or locations

oai_token-act-pair · gpt-3.5-turbo

terms related to desirability and undesirability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

kamp

-0.79

chin

-0.71

wrapper

-0.71

uania

-0.69

stock

-0.68

udeb

-0.67

 Anthem

-0.67

track

-0.67

packs

-0.66

brook

-0.66

POSITIVE LOGITS

 mosqu

1.07

 desirable

1.02

 undesirable

0.92

 distingu

0.90

 resil

0.88

 charact

0.82

igslist

0.77

 undermin

0.74

 behaviors

0.73

 adolesc

0.72

Activations Density 0.009%

words related to the desirability of objects, conditions, or locations

terms related to desirability and undesirability

No Comments

No Known Activations