INDEX

Explanations

phrases related to spatial direction and movement

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "backward" actions or movements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

raltar

-0.84

rament

-0.73

akings

-0.72

"},"

-0.71

riz

-0.71

AIN

-0.67

lez

-0.66

ozy

-0.66

aukee

-0.66

POSITIVE LOGITS

wards

0.95

ward

0.88

stairs

0.84

 backwards

0.82

 compat

0.79

step

0.79

gradation

0.78

dash

0.77

 backward

0.76

 spiral

0.72

Activations Density 0.005%

phrases related to spatial direction and movement

references to the concept of "backward" actions or movements

No Comments

No Known Activations