INDEX

Explanations

dates presented in a consistent format

oai_token-act-pair · gpt-3.5-turbo

dates and references to significant events

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

ktop

-0.65

\\\\\\\\

-0.59

çĶ°

-0.58

Ô

-0.57

esson

-0.56

Ö

-0.55

llan

-0.55

quit

-0.55

ivas

-0.54

è»

-0.52

POSITIVE LOGITS

 respective

0.78

 apiece

0.65

 respectively

0.59

including

0.58

 ranging

0.55

regular

0.55

 varying

0.55

 including

0.54

 individually

0.54

 interacted

0.54

Activations Density 2.347%

dates presented in a consistent format

dates and references to significant events

No Comments

No Known Activations