INDEX

Explanations

words related to launching or initiating actions or events

oai_token-act-pair · gpt-3.5-turbo

actions related to military aggression and attacks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Perspective

-0.63

 Tradable

-0.61

 representing

-0.59

ĻĤ

-0.59

adequ

-0.59

,,,,

-0.58

ilated

-0.58

 UCHIJ

-0.57

liness

-0.57

POSITIVE LOGITS

 missiles

1.08

 barrage

1.01

 volley

0.99

tir

0.93

 missile

0.92

 projectile

0.92

 rockets

0.86

 fury

0.86

 preempt

0.85

 torped

0.84

Activations Density 0.267%

words related to launching or initiating actions or events

actions related to military aggression and attacks

No Comments

No Known Activations