INDEX

Explanations

references to legal or financial terms, specifically related to damages and compensation

oai_token-act-pair · gpt-3.5-turbo

references to legal claims for damages

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bun

-0.68

 Garner

-0.67

 Positive

-0.67

 Handle

-0.66

 Purpose

-0.66

BOOK

-0.65

 Neigh

-0.63

 Merc

-0.63

 Ident

-0.63

OB

-0.61

POSITIVE LOGITS

 damages

1.25

 damage

0.96

damage

0.95

rays

0.84

ashtra

0.81

jury

0.79

hement

0.77

 mitigation

0.76

hiba

0.76

ionage

0.75

Activations Density 0.007%

references to legal or financial terms, specifically related to damages and compensation

references to legal claims for damages

No Comments

No Known Activations