INDEX

Explanations

phrases related to processes of ending or removing something

oai_token-act-pair · gpt-3.5-turbo

references to withdrawal in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

eric

-0.73

 ILCS

-0.69

seless

-0.69

nai

-0.68

ando

-0.68

xxxxxxxx

-0.67

ovie

-0.67

++++++++++++++++

-0.65

OSH

-0.63

 Borough

-0.63

POSITIVE LOGITS

 withdrawals

1.11

 withdrawal

1.02

 blockade

0.79

 gratification

0.77

 blockers

0.76

 Measures

0.76

avior

0.76

 threshold

0.75

ptives

0.75

 cessation

0.74

Activations Density 0.011%

phrases related to processes of ending or removing something

references to withdrawal in various contexts

No Comments

No Known Activations