INDEX

Explanations

names of places or people mentioned in the input text

oai_token-act-pair · gpt-3.5-turbo

frequent names and identifiers related to events or locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

soDeliveryDate

-0.76

ccording

-0.74

 eleph

-0.69

ãĥĺãĥ©

-0.66

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.65

accompan

-0.65

ãĤ¼ãĤ¦ãĤ¹

-0.63

 Princ

-0.62

 Parables

-0.61

;;;;;;;;;;;;

-0.60

POSITIVE LOGITS

ombat

0.84

online

0.67

ource

0.65

lore

0.62

ename

0.62

 Xander

0.61

xp

0.61

enum

0.59

apse

0.59

ikan

0.58

Activations Density 0.207%

names of places or people mentioned in the input text

frequent names and identifiers related to events or locations

No Comments

No Known Activations