INDEX

Explanations

terms related to conflicts or confrontations, such as "standoff" and "skirmish"

oai_token-act-pair · gpt-3.5-turbo

references to standoffs or conflicts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odor

-0.74

Cola

-0.74

GC

-0.73

living

-0.69

incarn

-0.67

undai

-0.65

rich

-0.65

gil

-0.65

Benef

-0.65

sample

-0.64

POSITIVE LOGITS

 standoff

1.42

terness

0.80

 siege

0.79

ishly

0.76

emate

0.75

 negotiators

0.74

ength

0.71

 SWAT

0.70

 confrontation

0.69

 showdown

0.69

Activations Density 0.010%

terms related to conflicts or confrontations, such as "standoff" and "skirmish"

references to standoffs or conflicts

No Comments

No Known Activations