INDEX

Explanations

phrases related to exerting control or influence

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dictate" and its variants, indicating control or authority

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

alo

-0.75

ovember

-0.74

oney

-0.72

Mania

-0.72

uni

-0.70

 Atomic

-0.69

ft

-0.67

mania

-0.66

 Cure

-0.65

groups

-0.65

POSITIVE LOGITS

 dictated

1.16

 dictate

1.13

 dict

1.10

 dictates

1.04

 confir

0.94

sheets

0.76

 govern

0.74

 curve

0.73

 lapt

0.71

 guiding

0.71

Activations Density 0.033%

phrases related to exerting control or influence

instances of the word "dictate" and its variants, indicating control or authority

No Comments

No Known Activations