INDEX

Explanations

words related to psychological states or emotions

oai_token-act-pair · gpt-3.5-turbo

terms related to themes of experimentation and exploitation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hips

-0.93

heny

-0.89

ento

-0.84

nings

-0.83

itu

-0.83

sen

-0.81

ingham

-0.77

iating

-0.77

IFE

-0.76

yss

-0.76

POSITIVE LOGITS

 grab

0.79

 ploy

0.77

 cipher

0.75

 hawk

0.73

 whore

0.73

 brake

0.73

 bunny

0.73

 glove

0.72

 hatch

0.72

 porn

0.72

Activations Density 0.301%

words related to psychological states or emotions

terms related to themes of experimentation and exploitation

No Comments

No Known Activations