INDEX

Explanations

words related to legal terms

oai_token-act-pair · gpt-3.5-turbo

terms related to specific names or identifiers that have an 'end' or similar structure

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

 lest

-0.70

 afore

-0.67

 fuzzy

-0.67

 Takeru

-0.67

 Pixar

-0.65

ghan

-0.62

 tackling

-0.62

 tempered

-0.62

bender

-0.61

swick

-0.60

POSITIVE LOGITS

bilt

1.12

imir

0.77

¼

0.76

¶æ

0.74

arious

0.74

rone

0.73

ĪĴ

0.73

Ĩ

0.72

alle

0.72

emort

0.72

Activations Density 0.083%

words related to legal terms

terms related to specific names or identifiers that have an 'end' or similar structure

No Comments

No Known Activations