INDEX

Explanations

references to the concept of someone being removed from a position of power or authority

oai_token-act-pair · gpt-3.5-turbo

terms associated with the concept of "deposition" or loss of power and influence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

nings

-0.77

 Heritage

-0.76

é¾

-0.75

 Pose

-0.75

hua

-0.75

 Hole

-0.74

ãĥīãĥ©ãĤ´ãĥ³

-0.73

Sense

-0.73

 Marble

-0.72

WHO

-0.72

POSITIVE LOGITS

dep

1.25

raved

1.02

orters

0.93

utations

0.92

utation

0.92

recated

0.88

ravity

0.85

ository

0.85

reci

0.83

orter

0.81

Activations Density 0.006%

references to the concept of someone being removed from a position of power or authority

terms associated with the concept of "deposition" or loss of power and influence

No Comments

No Known Activations