INDEX

Explanations

phrases related to requests, desires, and priorities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "the" and references to specific topics or requests for information

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

 namely

-0.92

igmatic

-0.79

ilon

-0.77

cial

-0.74

ategory

-0.74

aiden

-0.73

ornings

-0.73

 according

-0.73

adder

-0.72

abel

-0.71

POSITIVE LOGITS

 feds

1.20

 remainder

1.10

 rest

1.04

 next

1.02

 outcome

1.01

 courts

1.00

 Legislature

1.00

 government

0.98

 perpetrators

0.94

 entire

0.94

Activations Density 0.334%

phrases related to requests, desires, and priorities

occurrences of the word "the" and references to specific topics or requests for information

No Comments

No Known Activations