INDEX

Explanations

words related to negative sentiments or feelings, particularly strong dislike or hatred

oai_token-act-pair · gpt-3.5-turbo

instances of the term "loathe" or related expressions of strong dislike

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Norn

-0.84

rity

-0.81

*/(

-0.76

hower

-0.76

glass

-0.75

rition

-0.72

sonian

-0.71

lished

-0.69

ITAL

-0.67

manship

-0.67

POSITIVE LOGITS

oser

1.04

aning

1.02

aned

1.01

veland

1.00

aves

0.95

fty

0.92

zzle

0.92

ppy

0.91

ven

0.90

obb

0.90

Activations Density 0.009%

words related to negative sentiments or feelings, particularly strong dislike or hatred

instances of the term "loathe" or related expressions of strong dislike

No Comments

No Known Activations