INDEX

Explanations

phrases related to negative or concerning events or situations

oai_token-act-pair · gpt-3.5-turbo

expressions of concerns or negative situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

pel

-0.85

ritis

-0.81

á

-0.81

neys

-0.80

cellence

-0.80

arten

-0.79

rian

-0.78

rosis

-0.77

pei

-0.77

gerald

-0.76

POSITIVE LOGITS

 Revelations

0.98

 revelations

0.83

ly

0.78

 occurrences

0.78

Trend

0.77

 Trend

0.72

LY

0.72

 disclosures

0.71

ingly

0.71

 alarming

0.70

Activations Density 0.034%

phrases related to negative or concerning events or situations

expressions of concerns or negative situations

No Comments

No Known Activations