INDEX

Explanations

Medical or forensic terminology related aspects or actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "determined."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

oa

-0.70

AH

-0.68

era

-0.66

obi

-0.65

gate

-0.62

nen

-0.61

MY

-0.60

umbai

-0.60

ah

-0.59

 Quote

-0.59

POSITIVE LOGITS

 determined

3.59

etermined

1.98

 determination

1.93

 calculated

1.67

 determines

1.65

 determine

1.56

 determining

1.43

 determin

1.37

 measured

1.37

 decided

1.35

Activations Density 0.022%

Medical or forensic terminology related aspects or actions

instances of the word "determined."

No Comments

No Known Activations