INDEX

Explanations

words related to important or impactful terms or events

oai_token-act-pair · gpt-3.5-turbo

names and entities, particularly focusing on characters or individuals

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

¶ħ

-0.69

cffff

-0.61

berra

-0.61

taboola

-0.60

 ..."

-0.60

-0.59

''.

-0.55

 EDIT

-0.53

REL

-0.53

 interchange

-0.53

POSITIVE LOGITS

enges

0.63

 Franch

0.62

 sails

0.62

tsy

0.61

culosis

0.60

stood

0.60

doors

0.60

's

0.59

ians

0.57

hyde

0.57

Activations Density 0.372%

words related to important or impactful terms or events

names and entities, particularly focusing on characters or individuals

No Comments

No Known Activations