INDEX

Explanations

mentions or references to the front lines or engaging in direct action or confrontation

oai_token-act-pair · gpt-3.5-turbo

phrases related to "front line" themes and contexts

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

issions

-0.82

redited

-0.74

Pitt

-0.72

itt

-0.72

Arch

-0.71

Asset

-0.71

arth

-0.70

ission

-0.69

orig

-0.69

reditary

-0.69

POSITIVE LOGITS

 fighting

0.91

lines

0.79

 frontline

0.76

roads

0.73

fighters

0.72

breaker

0.71

 trenches

0.71

 combating

0.70

 lookout

0.70

 Assassins

0.68

Activations Density 0.033%

mentions or references to the front lines or engaging in direct action or confrontation

phrases related to "front line" themes and contexts

No Comments

No Known Activations