INDEX

Explanations

phrases indicating that something is subject to change

oai_token-act-pair · gpt-3.5-turbo

phrases indicating policies or conditions that are open to change or variability

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

resa

-0.75

ared

-0.71

ndra

-0.69

meal

-0.67

aghan

-0.67

started

-0.66

ergy

-0.66

rib

-0.66

dad

-0.66

 Crate

-0.65

POSITIVE LOGITS

 ridicule

0.96

 scrutiny

0.90

 stringent

0.85

 rigorous

0.85

 continual

0.84

 varying

0.81

 considerable

0.79

 constant

0.78

 extinction

0.78

 periodic

0.74

Activations Density 0.055%

phrases indicating that something is subject to change

phrases indicating policies or conditions that are open to change or variability

No Comments

No Known Activations