INDEX

Explanations

phrases indicating a conclusion or summary

oai_token-act-pair · gpt-3.5-turbo

phrases related to key conclusions or essential points

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

 UCHIJ

-0.81

atars

-0.76

ãĤ¼ãĤ¦ãĤ¹

-0.70

ald

-0.67

UCH

-0.67

ãĥ¼ãĥĨ

-0.66

MET

-0.63

lat

-0.63

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.62

ãĥ¡

-0.62

POSITIVE LOGITS

 margin

0.74

nings

0.73

apse

0.71

chart

0.69

aminer

0.69

inez

0.66

stakes

0.66

($)

0.63

claimer

0.63

margin

0.62

Activations Density 0.031%

phrases indicating a conclusion or summary

phrases related to key conclusions or essential points

No Comments

No Known Activations