INDEX

Explanations

terms related to political figures and actions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "ence."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĵĺ

-0.82

syn

-0.74

icable

-0.74

ding

-0.73

olicited

-0.72

STON

-0.71

acca

-0.70

ISO

-0.69

Spot

-0.68

ocular

-0.67

POSITIVE LOGITS

llor

0.95

ence

0.91

lihood

0.89

lement

0.85

phal

0.85

ment

0.80

ENCE

0.78

mble

0.77

rences

0.72

pend

0.70

Activations Density 0.009%

terms related to political figures and actions

occurrences of the word "ence."

No Comments

No Known Activations