INDEX

Explanations

mentions of legal terms and titles, especially related to legal recourse

oai_token-act-pair · gpt-3.5-turbo

specific names and terms associated with legal and bureaucratic contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

liest

-0.77

jay

-0.72

 Ãľ

-0.72

||||

-0.71

fol

-0.66

gers

-0.64

ãĥ³ãĤ¸

-0.64

cery

-0.63

gery

-0.63

jee

-0.61

POSITIVE LOGITS

theless

1.00

anwhile

0.84

amily

0.82

 DRAG

0.80

hovah

0.80

ember

0.78

untu

0.77

imately

0.77

perty

0.77

chio

0.76

Activations Density 0.029%

mentions of legal terms and titles, especially related to legal recourse

specific names and terms associated with legal and bureaucratic contexts

No Comments

No Known Activations