INDEX

Explanations

verbs related to causing potential negative impact, risk, or doubt

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "put" and its variations, indicating a focus on actions or consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

|--

-0.71

ificent

-0.69

cade

-0.67

HOW

-0.62

xxxx

-0.62

uary

-0.62

cue

-0.60

âĹ¼

-0.60

ocaly

-0.60

è¦ļéĨĴ

-0.59

POSITIVE LOGITS

 pressure

1.21

 undue

1.17

 strain

1.12

 emphasis

1.03

 brakes

1.03

 downward

0.99

 constraints

0.96

 limits

0.92

 strains

0.92

 forth

0.88

Activations Density 0.061%

verbs related to causing potential negative impact, risk, or doubt

occurrences of the word "put" and its variations, indicating a focus on actions or consequences

No Comments

No Known Activations