INDEX

Explanations

words related to ideas, thoughts, and feelings

oai_token-act-pair · gpt-3.5-turbo

expressions related to validation and insight

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

equ

-0.51

©¶æ

-0.49

interrupted

-0.42

appropriately

-0.40

 Powder

-0.40

eb

-0.39

inka

-0.38

Jav

-0.37

 Glory

-0.37

 equivalents

-0.37

POSITIVE LOGITS

byss

0.56

PsyNetMessage

0.54

Merit

0.47

IONS

0.46

 showc

0.45

ioch

0.44

 regarding

0.43

oried

0.43

 âĢº

0.43

gae

0.42

Activations Density 1.072%

words related to ideas, thoughts, and feelings

expressions related to validation and insight

No Comments

No Known Activations