INDEX

Explanations

phrases that indicate an action is being done with a specific purpose or goal in mind

oai_token-act-pair · gpt-3.5-turbo

sequential phrases that introduce or elaborate on information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

etheless

-0.71

ogly

-0.68

Laughs

-0.65

rider

-0.64

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.60

qus

-0.60

obook

-0.59

hedral

-0.59

minster

-0.58

arch

-0.58

POSITIVE LOGITS

 however

0.98

 please

0.97

 preferably

0.82

please

0.79

 suffice

0.74

 moreover

0.74

 multiply

0.73

 namely

0.72

 meanwhile

0.67

we

0.67

Activations Density 0.307%

phrases that indicate an action is being done with a specific purpose or goal in mind

sequential phrases that introduce or elaborate on information

No Comments

No Known Activations