INDEX

Explanations

terms related to acquiring or obtaining something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "obtain" and its variations in reference to acquiring permissions or items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hov

-0.79

roller

-0.76

heartedly

-0.75

psc

-0.70

gae

-0.69

edo

-0.67

cas

-0.64

evil

-0.62

quartered

-0.61

lington

-0.61

POSITIVE LOGITS

 permission

1.07

hold

0.87

 access

0.82

 signatures

0.81

 information

0.79

 firsthand

0.79

 licenses

0.78

 copies

0.77

 unrestricted

0.76

 authorization

0.76

Activations Density 0.081%

terms related to acquiring or obtaining something

instances of the word "obtain" and its variations in reference to acquiring permissions or items

No Comments

No Known Activations