INDEX

Explanations

phrases related to being in a position of leadership or influence

oai_token-act-pair · gpt-3.5-turbo

instances of the word "led" indicating leadership or causation in various contexts

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

 sunset

-0.77

 Scotia

-0.64

endi

-0.59

cube

-0.59

Done

-0.58

 Smoking

-0.58

afort

-0.57

 Ratio

-0.57

 Buildings

-0.56

 symmetry

-0.56

POSITIVE LOGITS

gers

1.22

gling

1.00

ges

0.93

better

0.92

ging

0.90

gered

0.87

ger

0.85

ged

0.83

leaders

0.82

lead

0.79

Activations Density 0.029%

phrases related to being in a position of leadership or influence

instances of the word "led" indicating leadership or causation in various contexts

No Comments

No Known Activations