INDEX

Explanations

words related to minimizing, reducing, or preventing

oai_token-act-pair · gpt-3.5-turbo

actions or concepts associated with reducing negative impacts or minimizing risks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ker

-0.87

enegger

-0.85

king

-0.81

swick

-0.76

otle

-0.75

gob

-0.74

join

-0.73

cart

-0.71

leader

-0.69

worldly

-0.67

POSITIVE LOGITS

imize

0.92

 minimizing

0.82

 minimize

0.80

 distractions

0.78

 amounts

0.77

 maximizing

0.77

 minimized

0.71

 misunderstand

0.71

 utilization

0.71

imal

0.71

Activations Density 0.034%

words related to minimizing, reducing, or preventing

actions or concepts associated with reducing negative impacts or minimizing risks

No Comments

No Known Activations