INDEX

Explanations

phrases related to guidance, instructions, and following a specific path or set of rules

oai_token-act-pair · gpt-3.5-turbo

phrases related to following rules and instructions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

tu

-0.80

etheless

-0.69

venge

-0.68

vu

-0.67

enf

-0.66

cest

-0.65

afety

-0.64

omnia

-0.64

pora

-0.63

adjusted

-0.63

POSITIVE LOGITS

 footsteps

1.19

 closely

1.02

 steps

0.92

 instructions

0.91

 path

0.85

 directions

0.81

 route

0.81

 blindly

0.78

 guidelines

0.75

 whims

0.73

Activations Density 0.178%

phrases related to guidance, instructions, and following a specific path or set of rules

phrases related to following rules and instructions

No Comments

No Known Activations